Fraud is a general
problem facing many sectors of commerce and industry. The FEDAURA
Project is aimed at taking forward novel technology from academic research
to show how it can address the problem of fraud in financial claims processing.
The project will allow the full commercial potential of the University research
to be realised. For demonstration purposes, the work is focused on the Benefit
Claims sector. However, the commercial partners will evaluate the methods
for use in other areas of business, based on the results gained from the project.
Fraudulent Benefit claims and clerical errors are estimated to cost UK taxpayers
between £2.5 and £7 billion per year, within a total benefit bill
of around £100 billion per year [BA]. The major objective
in the FEDAURA Project is to find methods that
help identify fraudulent benefit claims, thus lowering the cost of fraud to
the UK. It aims to demonstrate that the AURA technology under development
at University of York can be readily adapted and scaled to address such real-world
problems. The potential savings for the UK are considerable, and the effect
of nullifying only a modest proportion of frauds could have a significant
impact for the UK. The DWP aims to reduce benefit losses from fraud
and error in Income support (IS) and Job Seeker's Allowance (JSA) by 25% by
2004 and by at least 50% by 2006. DWP believe that the project has
the potential for making savings in the levels of fraud and error and directly
contributing to the achievement of the target.
[BA] BA internal Command paper, March 1999, A New Contract
for Welfare: Safeguarding Social Security.
Despite the potential benefits, the scope for improvement is limited by the
performance of currently available technology. New, powerful methods of pattern
detection are now available within University research that can be applied
to this problem with many potential advantages. In particular, an extension
into non-linear and probabilistic classification methods may bring large savings
for the DWP. This project will assess and develop these new methods
for application within the problem of fraud.
The project will allow the development of AURA technology in new application
areas, with knowledge of its application to be transferred to leading software
consultancy companies. It will also allow an SME to develop the technology
ready for market. In addition, the partners EDS, Sema and DWP
have no experience of AURA based methods, and the project will disseminate
the utility of these methods in comparison with other neural methods to them.
There is an element of risk in the project since the technology developed
may not perform as well as expected, and the commercial potential may be substantially
reduced, or even negligible, in that event. This risk aspect means that the
FEDAURA Project described here could not take
place without additional support from the DTI MI Programme.